Web Posted on: August 4, 1998

TetraNauta: A Wheelchair Controller
for Users with Very Severe
Mobility Restrictions

Antón Civit Balcells
Robotics and Computer Technology
for Rehabilitation Laboratory
Facultad de Informática
Universidad de Sevilla
Avda. Reina Mercedes s/n.
E-41012 Sevilla. Spain
Tel.: + 34 5 455 27 79
fax: + 34 5 455 27 59
e-mail: civit@icaro.fie.us.es

Julio Abascal González
Laboratory of Human-Computer
Interaction for Special Needs
Informatika Fakultatea
Euskal Herriko Unibertsitatea
649 Postakutxa
E-20080 Donostia. Spain
Tel: + 34 43 448067
fax: + 34 43 219306
e-mail: julio@si.ehu.es

Summary

TetraNauta is an on-going R&D project aimed to develop a controller for standard electric-powered wheelchairs that permits users with very severe mobility restrictions (such as people with tetraplegy) to easily navigate in closed environments (home, hospital, school, etc.). This project intends to design a non-expensive guidance system to help this kind of users to drive the wheelchair with the minimum effort, but maintaining the user as active as possible.

This project, started in December 1996, was scheduled for three years. It is part of the Spanish National Rehabilitation Technology Project (PITER). It is co-ordinated by the National Hospital of Paraplegics from Toledo, with Bioingeniería Aragonesa as the industrial partner, and the Universities of Seville and the Basque Country in charge of the research.

| Top |  

Introduction

The benefits of autonomous mobility for physical and cognitive rehabilitation have frequently been stated. To advance in this way, diverse auto-guided "intelligent" wheelchairs have been designed for people having serious restrictions to use standard electric-powered wheelchairs. Most of these wheelchairs are able to move around avoiding obstacles and traversing difficult zones (such as narrow doors) (see [Bourhis-96], [Yoder-96], [Cooper-95], [Craig-93]). Nevertheless, from our experience designing an intelligent wheelchair ([Civit-97], [Díaz-97], [Civit-96]), some problems arise: On the one hand, to convert a standard electric-powered wheelchair into an "intelligent" one, internal parts have to be manipulated, resulting in the cancellation of the warranty. In addition, the high price of the resulting devices make this kind of wheelchair not affordable for the majority of the possible users. On the other hand, the use of smart wheelchairs may interfere the rehabilitation process, due to the lack of participation of the user, limiting the benefits that can be obtained if the remaining abilities are used.

To avoid these problems, the following requirements were defined for TetraNauta project:

These guidelines have been followed in the implementation of the TetraNauta prototype: landmark detection has been used in the automatic driving mode (better than environment recognition) to reduce complexity and price. The system is connected to standard buses to avoid internal modifications and a new adaptable interface has been developed.

The project started with a user requirement study carried out by the National Hospital of Paraplegics. As a part of this analysis a specific methodology for the detection and evaluation of the mobility needs of severely motor impaired people was developed ([NatHosp-98]). This methodology is mainly based on direct analysis of the actions and movements performed by the users. A market survey to determine the commercial viability of the product has been also conducted by the industrial partner ([Bioing-98]).

| Top |

Features of the TetraNauta prototype

The idea behind TetraNauta is to help users that have very strong difficulties in riding a chair by any of the conventional means. They have just to indicate where they want to go and the navigation is done automatically by the chair controller. The user will be able to recover manual control whenever he or she desires to do so.

Since full automatic guidance results too complex and expensive, the TetraNauta approach uses landmark detection techniques, similar to the ones frequently used by Automated Guided Vehicles. This technique has also some disadvantages: on the one hand wheelchairs can only follow the limited drivepath, so they are bounded to in-door medium. On the other hand, the environment must be adapted using detectable marks ([Arkin-90]). Nevertheless, these disadvantages are highly compensated by the low price of the used detectors (a standard black and white video camera). Currently, landmarks are lines painted in the floor, but different alternatives are being under study. The absolute position detection is solved using transponders.

TetraNauta architecture

The architecture of TetraNauta reflects the modular structure shown in fig. 1. Four main intercommunicating modules have been defined to manage external events and communications: User Interface, Manager of External Signals, Traffic Module and Power Stage Control. While other two modules, called Trajectory Planner and Trajectory Follower, manage the general co-ordination and supervision.

Different scenarios have been foreseen (big hospital, medium school, small day-centre, home), so the functionality of these modules may vary from one scenario to other. Lets briefly describe the functionality of each module.
 

User interface

The user controls the system through a very intuitive graphical interface. The interface translates his or her orders (e.g. the desired destination) into commands for the Trajectory Planner. It also gives feed-back to the user about the current operation. Diverse input/output devices are possible to suit the user physical characteristics: joystick, mouth-stick, touch-screen, one-key scanning, and voice recognition for input purposes, while a display and synthetic voice are used for output. The selection process has been designed to minimise the effort required from the user.
 

Trajectory Planner

This module receives a command from the user through the User Interface and calculates an optimised trajectory (in terms of distance and traffic) to the desired final location, using an internal representation of the environment in which the wheelchair is currently navigating. This trajectory is partitioned in segments and passed to the Trajectory Follower. This module is responsible for the supervision of the entire displacement. It can recalculate the trajectory if it receives new path restrictions from the trajectory follower or the traffic module.

Traffic module

If the number of wheelchairs is relatively high, traffic problems can arise, mainly in crossings. To avoid traffic jams, some scenarios may require traffic planning. Different models have been studied:

In the first case the traffic modules communicates with the external computer to send the desired trajectory and to receive information about the restrictions to this trajectory imposed by a central Traffic Manager. In the other two cases, there are not external restrictions and the traffic decisions are taken by each Traffic Module ([Alami-97], [Brummit-96]). Two approaches are studied. In the first one wheelchairs can communicate with the nearest ones by low range media (radio or infrared beams). In the second one, each wheelchairs behaves as an autonomous entity with a different behaviour ([Chaib-Draa-94]).

Traffic Controller

This module is only present in very crowded environments. It records, from the Traffic Modules, all the movements of the active wheelchairs. When traffic jams are detected or drivepaths are blocked or broken, the Traffic Controller sends this information to the active wheelchairs in form of restrictions to the map stored by each wheelchair. So the wheelchairs can recalculate a new trajectory to their final destination.
 

Trajectory Follower

External signals manager

The signals from the diverse "sensors" (infrared and ultrasonic detectors, transponders, video camera, etc.) and their programming are under the responsibility the External signals Manager, that communicates every event to the Trajectory Follower.
 

Power Stage Controller

This module produces the signals that actually control the motors. In the current demonstrators, the power stage interfaces with the standard wheelchair controller through the DXbus ([MEAD-97]) or analog signal. The possibility of the use of other standard buses, as the M3S ([M3S-97]), is currently studied. TetraNauta controllers are based on an embedded PC architecture. This has many advantages from the hardware interface and software development points of view but the economics of this approach for a future commercial version are still under study.

| Top |

Other issues

Safety

Safety problems are solved in the previous modules:

User participation

| Top |

Conclusions

Preliminary results in the different modules have been obtained, that demonstrate the technical feasibility of the system. To test the fulfilling of the planned requirements two important studies must be carried out:

These aspects require a high effort but they are essential to produce something really useful. So, obtaining resources for these purposes is currently the primary aim of the consortium.

| Top |

References:

| Top | | TIDE 98 Papers |